Disrupted functional connectivity between sub-regions in the sensorimotor areas and cortex in migraine without aura

Forty-eight patients with MWoA (mean ± SD age = 38.1 ± 10.4 years) and 48 healthy controls (HCs) (mean ± SD age = 39.0 ± 11.0 years) were enrolled from the outpatient clinic of the Department of Neurology at Shanghai Ninth People’s Hospital. Patients were diagnosed with MWoA by a neurologist based on the International Classification of Headache Disorders 3rd edition criteria [2]. Demographic and clinical data were registered and evaluated in our headache database, including age, sex, disease duration, attack frequency (times/month), attack duration (hours) and pain intensity of migraine attacks assessed by a visual analogue scale (VAS) [15]. Patients also completed the MIDAS [16] and HIT-6 [17] regarding the accurate assessment of their headache-related disability. All patients with MWoA were scanned during an interictal period, with no headache 48 h before or 24 h after MRI scans, and they did not report a migraine attack or discomfort during the MRI scans. Migraineurs reported that they did not take preventive medication and did not suffer chronic migraine. The HCs had no headaches or chronic pain disorders in the past year. Moreover, the immediate family members of the HCs did not suffer from migraine or other headaches. Other exclusion criteria were the following: left-handedness, drug abuse, any neurological or psychiatric diseases, metabolic diseases (e.g., diabetes mellitus) or cardiovascular diseases based on clinical examination and structured interviews. The demographic and clinical data are provided in Table 1.

MRI was carried out in a 3.0 Tesla Siemens Trio Tim system with a 12-channel head coil at the Shanghai Key Laboratory of Magnetic Resonance (East China Normal University, Shanghai, China). All subjects’ head movements were minimized with a Siemens dedicated filler. T1-weighted anatomical images were collected using a 3-dimensional magnetization-prepared rapid-acquisition gradient-echo pulse sequence (repetition time = 2530 ms, echo time = 2.34 ms, inversion time = 1100 ms, flip angle = 7°, number of slices = 192, sagittal orientation, field of view = 256 × 256 mm², matrix size = 256 × 256, slice thickness = 1 mm, 50% gap). Following this, subjects were asked to remain motionless, keep their eyes closed, and stay awake and relaxed. Resting-state images consisted of 210 volumes of a T2*-weighted gradient-echo echo planar imaging pulse sequence (33 slices of 3.5 mm thickness, repetition time = 2000 ms, echo time = 30 ms, flip angle = 90°, transverse orientation, field of view = 220× 220 mm², matrix size = 64 × 64, 25% distance factor).

Functional data were preprocessed using the Data Processing Assistant for Resting-State fMRI (DPARSF, http://​rfmri.​org/​DPARSF) and analyzed with Statistical Parametric Mapping (SPM12, http://​www.​fil.​ion.​ucl.​ac.​uk/​spm). To avoid scanner instability, the first 10 volumes were discarded. The images for each subject were motion-corrected, and in no subjects was there significant movement (> 2 mm in any direction). The data from all the subjects were used for the subsequent analysis. We utilized the Friston [18] 24-parameter model to regress out the head motion effects from realignment. Functional volumes were slice-timing corrected and coregistered to the individual structural images without resampling. The transformed structural images were segmented [19] and normalized to the Montreal Neurological Institute (MNI) space. Based on the same deformation field, functional images were then normalized to the MNI space (3 × 3 × 3 mm³) using the Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra (DARTEL) tool [20]. Spatial smoothing was performed on the functional images by a 6 mm FWHM Gaussian filter. We performed temporal bandpass filtering (0.01 < f < 0.1 Hz), and the mean signals from the white matter, cerebrospinal fluid, and linear trends were also included as covariates to reduce the influence of low-frequency drift and high-frequency respiratory and cardiac effects.

The mesial premotor cortex including pre-supplementary motor area (preSMA) and supplementary motor area proper (SMA), lateral premotor cortex including dorsal premotor cortex (PMd) and ventral premotor cortex (PMv), and primary sensorimotor cortex including primary motor cortex (M1) and primary somatosensory cortex (S1) have been identified as key cortical areas for sensorimotor function [14].The HMAT ROIs included M1, S1, the SMA, the preSMA, the PMd, and the PMv. We calculated Pearson’s correlation coefficients of the mean time courses extracted from the HMAT ROIs with the signals of whole-brain voxels. To improve the normality of the data distribution, we converted FC maps to FC Z-value maps by Fisher’s Z-transformation. Then, the voxelwise two-sample t-test with sex and age as covariates was conducted within a brain mask. The two-tailed (voxel-level) threshold was p < 0.001 in combination with familywise error (FWE) corrected to p < 0.05 at cluster level. Surviving clusters were deemed significant. We performed a normal distribution test and found that the clinical data did not conform to the normal distribution, so we used Spearman correlation test. Then individual mean z-values for the surviving clusters were extracted to determine Spearman’s correlation correlations with clinical data features, including disease duration, attack frequency, and attack duration, as well as VAS, MIDAS and HIT-6 scores. Thresholds were p < 0.05, one-tailed, and FDR corrected for multiple comparisons.

The demographic and clinical data of the migraine and control groups are presented in Table 1. The ages and gender proportions did not differ between the two groups.

The patients with MWoAs showed weaker FCs between L M1 and the right middle occipital gyrus, right middle temporal gyrus, right superior temporal gyrus, left superior parietal gyrus, right postcentral gyrus, and right middle frontal gyrus (Table 2, Fig. 1). We also found decreased FCs between R M1 and the right middle occipital gyrus, right superior temporal gyrus, left superior parietal gyrus, right postcentral gyrus, and right cuneus in patients with MWoAs relative to HC (Table 2, Fig. 1).

The patients with MWoAs showed weaker FCs between L S1 and the bilateral precuneus, bilateral postcentral gyrus, bilateral paracentral lobule, bilateral middle temporal gyrus, right superior temporal gyrus, left inferior parietal lobule and right inferior frontal gyrus and right middle frontal gyrus (Table 2, Fig. 1). We also found decreased FCs between R S1 and the bilateral precuneus, right cingulate gyrus, bilateral postcentral gyrus, bilateral paracentral lobule, bilateral precentral gyrus, right superior temporal gyrus, left inferior parietal lobe, and left occipital superior gyrus (Table 2, Fig. 1).

The patients with MWoAs showed weaker FCs between the L SMA and bilateral precentral gyrus, bilateral postcentral gyrus, left precuneus and bilateral inferior parietal lobule and right middle frontal gyrus, right cuneus, bilateral middle temporal gyrus, and right superior temporal gyrus (Table 2, Fig. 2). The patients with MWoAs showed weaker FCs between the R SMA and the right middle frontal gyrus, bilateral postcentral gyrus, bilateral inferior parietal lobule, bilateral cuneus, bilateral precuneus, and right middle temporal gyrus (Table 2, Fig. 2).

The patients with MWoAs showed weaker FCs between the L PMv and the left precentral gyrus, right cingulate gyrus, and right postcentral gyrus (Table 2, Fig. 2). The patients with MWoAs showed weaker FCs between the R PMv and the bilateral precuneus, bilateral postcentral gyrus and L SMA (Table 2, Fig. 2).

The patients with MWoAs showed a weaker FC between the R preSMA and the left precuneus (Table 2, Fig. 3).

The patients with MWoAs showed weaker FCs between the R PMd and the left precuneus, bilateral middle occipital gyrus, cerebellum areas, bilateral cuneus, and right postcentral gyrus (Table 2, Fig. 4).

A negative correlation was found between the disease duration of the patients and the FC Z scores between the L M1 and the R middle temporal gyrus (MTG). The pain intensity of the patients was negatively correlated with the FC Z scores between the L SI and the R superior temporal gyrus (STG). The FC Z scores between the L S1 and the R MTG and R STG were also negatively correlated with disease duration in MWoAs. The HIT-6 score in MWoAs was negatively correlated with the FC Z score between the L PMv and the R cingulate gyrus (Fig. 5).

Our previous study found that MWoAs exhibited functional abnormalities in the S1 and PMC and weaker FCs between the S1 and brain areas within the pain intensity and spatial regions involved in pain processing. In addition, we subdivided the sensorimotor areas as preSMA, SMA, PMd, PMv, M1, and S1 and found disrupted FCs between these sensorimotor brain regions and numerous other regions of the brain, such as the visual cortex, temporal cortex, posterior parietal lobule, prefrontal areas, precuneus, cingulate gyrus, sensorimotor areas proper and cerebellum areas. And we didn’t find any hyper sensorimotor connectivity in the MWoAs patients. These abnormal interactions between sensorimotor regions and multiple brain areas might contribute to dysfunction of the integration of information within and across the sensory modalities and, thus, could lead to the initiation of migraine attack and/or its associated symptoms. These impaired FCs were negatively correlated with clinical data, such as disease duration, pain intensity, and HIT-6 scores, which provided increased evidence that impaired FC between sub-regions of sensorimotor regions and multiple brain areas is involved in the pathophysiological mechanism of migraine.

M1 is the main contributor to generating neural impulses that pass down to the spinal cord and control the execution of movement [21] and has been implicated in the ascending trigemino-thalamo-cortical nociceptive pathway [8]. Activation of S1 in studies of laser-evoked pain supports a role for S1 in the sensory aspects of pain, including discrimination and localization of pain intensity [22]. The SMA has many proposed functions, including the internally generated planning of movement, the planning of sequences of movement, and the coordination of the two sides of the body, such as in bi-manual coordination [23, 24]. The SMA is involved in pain anticipation [25]. The activity of pre-SMA neurons suggests that it involves a mechanism for switching from automatic to controlled eye movements [26]. The PMd may participate in learning to associate arbitrary sensory stimuli with specific movements or learning arbitrary response rules [27, 28]. The PMv is often studied with respect to its role in the sensory guidance of movement. Neurons here are responsive to tactile stimuli, visual stimuli, and auditory stimuli [29‐31].

We found decreased functional connectivity between M1 and many other brain areas, such as the right middle occipital gyrus, right cuneus, right middle and superior temporal gyrus, left superior parietal gyrus, right postcentral gyrus, and right middle frontal gyrus, which are involved in pain perception and pain processing. A meta-analysis showed that excitatory M1 stimulation had a significant effect on reducing headache intensity and frequency of headache attacks in patients with migraine with a large effect size [11]. The middle occipital gyrus and right cuneus are involved in visual processing, and the migraineurs showed significantly higher activation than the control group in the left cuneus while viewing negative pictures [32]. The temporal pole participates in pain processing by mediating affective responses to painful stimuli and by acting as a multisensory integration zone responsible for processing painful, visual, auditory, and olfactory stimuli [3, 33, 34]. The posterior parietal lobule is now believed to underpin higher-order processes of sensory inputs, multisensory and sensorimotor integration, spatial attention, intention, and the conjoint representation of external space and the body [35]. The somatosensory cortex is contained in the postcentral gyrus, and a previous study found functional abnormalities in S1 and disrupted functional connectivity between S1 and other cortical regions [4]. The MFG, as part of the prefrontal cortex, is thought to be involved in the cognitive evaluation and modulation of pain [36]. The frequency of migraine attacks and the duration of the disorder had a significant impact on cortical thickness in the sensorimotor cortex and the middle frontal gyrus [37]. Disrupted FC between M1 and these brain regions might contribute to visual processing, multisensory integration, spatial attention and intention abnormalities, and dysfunction in the cognitive evaluation and modulation of pain.

We found decreased functional connectivity between S1, the SMA and many other brain areas, such as the bilateral precuneus, bilateral postcentral gyrus, bilateral paracentral lobule, bilateral precentral gyrus, right cingulate gyrus, bilateral middle temporal gyrus, right superior temporal gyrus, bilateral inferior parietal lobule, right inferior and middle frontal gyrus, and left occipital superior gyrus and right cuneus. In addition, patients with MWoAs showed a weaker FC between the R preSMA and the left precuneus. The precuneus has been proposed to participate in information transfer and multimodal integration, which might be essential for the processing of spontaneous thoughts and for internal awareness [38]. The postcentral gyrus, paracentral lobule, and precentral gyrus are involved in sensorimotor networks, and a previous study found functional abnormalities and abnormal cortical thickness or gray matter volume in the sensorimotor network in migraineurs [4, 39, 40]. Gray matter volume decreases in the right anterior cingulate are related to the estimated frequency of headache attacks [41]. The temporal pole has been demonstrated to participate in pain processing and multisensory integration [3, 33, 34]. The inferior parietal lobule has been demonstrated to be involved in spatial discrimination and attention to pain [42‐44]. The prefrontal region is thought to be involved in the cognitive evaluation and modulation of pain [36]. The middle occipital gyrus and right cuneus are involved in visual processing, and significantly increased cortical thickness was found in the lateral occipital cortex [45].

Brain areas with decreased functional connection to PMC.

The bilateral PMCs were activated during a delayed match-to-sample task using thermal stimuli in healthy people; they receive input information from the ACC and contribute to the communication and selection of a decision about the nature of the afferent sensory information during both intensity and spatial discrimination [5]. The patients with MWoAs showed weaker FCs between R PMd and the left precuneus, bilateral middle occipital gyrus, cerebellum regions, bilateral cuneus, and right postcentral gyrus. The patients with MWoAs showed weaker FCs between the PMv and left precentral gyrus, right cingulate gyrus, bilateral precuneus, bilateral postcentral gyrus and L SMA. Numerous regions mentioned above are involved in the pathology of migraine. In addition, a previous study found structural changes in cerebellum-associated migraine [46]. The cerebellum has been demonstrated to be involved in human nociception [47] and is even suggested to play a modulating role in pain perception.

The disease duration of the patients was negatively correlated with the FC Z scores between the L M1 and the R MTG and R STG and the FC Z scores between the L S1 and the R MTG and R STG, which suggests that recurrent headache attacks may drive functional changes and contribute to this disrupted network among the L M1, L S1 and temporal regions. The pain intensity of the patients was negatively correlated with the FC Z scores between the L SI and R STG. This result further indicates that the reduced FC between the L S1 and R STG is indeed a maladaptive functional plasticity closely related to pain intensity in MWoAs. The HIT-6 score in MWoAs was negatively correlated with the FC Z score between the L PMv and R cingulate gyrus, which suggests that the reduced FC is significantly related to the impact of recurrent migraine attacks on daily life.

Although our research revealed that MWoAs exhibited deficits in the functional connectivity between sensorimotor regions and other cortical areas, the current study had several limitations. First, we examined patients only in the interictal phase, and therefore, functional connectivity in the ictal phase must also be explored. Second, we focused on MWoAs but did not examine migraineurs with aura and chronic migraineurs. In a future study, we will evaluate the functional abnormalities in the sensorimotor regions of these other two populations during the ictal phase.